Tools you can use
- By Brian Robinson
- Sep 30, 2002
In the past year's discussions about the use of technology to boost homeland security efforts, one consistently understated issue is the usability of that technology — the so-called human factor. After all, if the technology ends up hindering rather than helping people, what use is it?
It's a point that the National Academies makes toward the end of its recent report, "Making the Nation Safer: The Role of Science and Technology in Countering Terrorism."
"The ease and effectiveness of use of information and other technical outputs by the people they are intended to support" is a key aspect in the deployment of any technology, the report says. Design and deployment of the technology systems "must take human, social and organizational factors into account."
Providers of technology systems have been aware of this, but in the past, they felt they had to make ease of use a secondary consideration to the immediate needs of the market, experts say.
Geographic information systems, for example, are expected to play an increasing role in disaster management. But if the director of a disaster management center is not GIS-savvy, he will need to find a GIS expert to produce and explain the information he needs — opening the situation to delay and misunderstanding.
"I find myself considering the fact that we in the GIS industry often seem to be playing with things that are technically neat, but that are overwhelming to the users," said Russ Johnson, director of public safety for GIS vendor ESRI.
Researchers are trying to keep usability in mind as they look for better ways to connect powerful computers and systems with the abilities of the people who have to use them.
Thinking Like a Human
Researchers' focus on usability results from a sense that computer design has typically favored the needs of the computer, or at least what computer engineers have traditionally considered important. Most systems and applications now use flashing icons and brightly colored interfaces to display computer-generated data. Although some people would consider such features to be user-friendly, most researchers deem them insufficient for true understanding and interaction.
The bottom line: Researchers are putting the human back into the equation.
"We as technologists tend to analyze how things are done currently, and then we take that process and try to substantiate it in a digital system," said Ward Page, a program manager at the Defense Advanced Research Projects Agency. "But that's the wrong approach. When we move away from the visual approach of such things as paper maps we lose resolution, and that's important for humans."
Software experts need to re-examine this approach and question why they do what they do with technology, he said.
One of the projects at Sandia National Laboratories, for example, is trying to provide a computer with the sensibilities of a human so that it can eventually work with its human operators more like a person would.
"When people work with each other over a length of time, they come to have a rich knowledge of the other person and how they think and react in certain circumstances," said Chris Forsythe, a project leader in Sandia's computational initiatives group. "People use that knowledge to create much more effective interactions than they would otherwise be able to have."
Unlike a system based on artificial intelligence, a common approach that uses defined rules to capture expert knowledge, the Sandia project is striving to create a computer model that more closely imitates the cognitive way that humans operate — by using patterns and cues that enable them to recognize familiar situations and react accordingly.
Such a model could be employed for specific situations, Forsythe said, such as ones that approximate how a counterintelligence agent, a security analyst, a forensic technologist or a security guard might react to certain information. One system working alone, or several together, might be used to identify suspicious trends in a set of data.
Simplicity is Best
At the Idaho National Engineering and Environmental Laboratory
(INEEL), researchers focus more on traditional support systems, emphasizing simplicity and power so that systems that provide high-level modeling and predictive capabilities can be used by people with differing skill levels.
RapidOps, which will likely go into beta testing in October or November, is a Microsoft Corp. Windows-based time-and-probability modeling tool that can be used for assessing such things as physical security vulnerabilities and necessary responses. It creates metaphors that are easier to relate to, such as pitting one virtual team against another in given situations and predicting the probabilities of one team's chances of winning.
One goal is to develop a way to model things in four dimensions — with time as the fourth dimension — represented in an integrated, intuitive way so it is easier for people to absorb and understand. However, it is important to ensure that technologies are helpful without being distracting, said Jerry Harbour, a consulting scientist at INEEL.
"The challenge is to find technology that truly is a support for people and that doesn't overwhelm them," he said. "I see systems all of the time that are very gadget-laden, but people use maybe 10 percent of the features at best."
RapidOps is designed to offload, or to transfer work from one computer to another, as many tasks as possible, and the graphics-based interface is focused on tasks the human user can do quickly. The idea is to show users that they are dealing with a transparent, intuitive system and that they can trust it to help them.
"It should take no longer than 10 or 15 minutes to learn," Harbour said. "Our goal in the beta phase is get a 'Wow!' response from people so they will readily accept it for infusion into the current situations they are dealing with."
Using all the Senses
The overarching aim of most of this research is to adapt the output of computer systems and applications to better fit the way people use their senses and abilities when reacting to a situation. Keyboards, graphics pads and monitors are the most direct ways to transfer data to and from computers, but they are not the best ways for people.
Researchers believe that more multimodal methods of manipulating and analyzing data, which seamlessly incorporate such things as sight, sound, touch and voice, have to become the norm in order to use technology more effectively in crisis situations.
Most work in the GIS field, for example, has been focused on how to optimize access to the GIS database rather than analysis of the data itself, said Alan MacEachren, director of Pennsylvania State University's GeoVISTA center.
However, as the use of GIS in more mainstream settings increases, it needs to become more user-friendly. In addition, there's a growing realization that GIS is often used by groups of people, not individuals, so the technology also must address group dynamics.
"All GIS technologies have been designed with one operator at a time in mind," MacEachren said. "The focus now is on building tools that people can use to collaborate in real time and also to be able to collaborate asynchronously in different time periods."
One GeoVISTA project, called DAVE-G for Dialogue-Assisted Visual Environment for GeoInformation, is trying to make use of the way that people use speech and gesture when communicating ideas.
For example, several people in an emergency center could be standing in front of a large regional map on a wall and could isolate particular problems by gesturing to areas on the map while talking, with the display changing accordingly. The final result would ideally be a resolution of their discussion accurately represented visually on the map.
But researchers need a better understanding of how humans use their senses to derive knowledge, MacEachren said. Basic technologies such as speech recognition need to be improved and made to work with spatial information systems.
Leaers of the GIS industry understand the need for a multimodal approach to using geospatial data, said Matt Tate, director of the federal business unit at Intergraph Corp.'s Mapping and GIS Solutions. The problem is how to include that in products that customers will want to buy.
"Users require that the ways they operate don't change too much and that they don't have to change their workflow," he said. "In the past, it was definitely all about which tools were the coolest, which had the most pizzazz. Now you almost have to 'underwhelm' them."
However, DARPA's Page believes that most people recognize that interface improvements are needed, and once they understand the trade-offs that go along with possible changes, they will readily accept the new ways of working.
Robinson is a freelance journalist based in Portland, Ore. He can be reached at [email protected]